Alzheimer's disease is a progressive disorder of unknown etiology and without a definitive in vivo diagnosis. Unfortunately, relatively few CT or PET studies have examined within patient (longitudinal) changes. Several longitudinal CT studies including our own suggest that ventricular and cortical changes are of little value in predicting decline. However, among patients that show clinical deterioration nearly all show progressive temporal lobe atrophy. Cross sectional PET studies have consistently revealed both global and regional (particularly temporo-parietal) reductions in glucose utilization. However, it is unknown if these studies are of predictive value in AD. Our pilot data suggests that there may be some predictive value in early cases. Our recent PET work has identified an AD subgroup that is identified by periventricular CT white matter lucencies (PWML), gait dysfunctions, and preserved metabolic rates. It is of great interest that normals with PWML show metabolic reductions, ventricular enlargement and evidence for motor slowing. With the proposed longitudinal study, we anticipate increasing our understandings of the relationships between metabolic changes and the clinical course of the 2 PWML groups. Further, we shall examine in AD and PWML the longitudinal metabolic significance of structural brain lesions. We propose 2 related studies. Study 1 will be a longitudinal study of 48 AD and 32 control subjects; half of each group will have CT evidence of PWML. Using repeated-measures multivariate analyses of variance and correlation procedures, we intent to investigate regional metabolic rates and their association with clinical deterioration and white matter disease. Study 2 will examine 80 questionably memory impaired subjects with and without PWML. Based on our recent behavioral findings, 15% of this group can be predicted to show further cognitive decline over 2 years. We hypothesize that the patients with white matter changes, lower metabolic rates and increased ventricle size will show declines. Overall, these studies offer the possibility of identifying the earliest brain changes and the specific correlates of progressive dementia in AD.